Performance Evaluation of Data Rate Adaptation Mechanisms for LoRaWAN networks for Scenarios of Livestock in Semi-Confinement
LoRaWAN, ADR, Machine Learning, IoT.
This work aims to investigate Adaptive Data Rate (ADR) mechanisms in LoRaWAN networks as a solution for dynamic IoT scenarios and also propose a new solution based on latter investigations. The standard ADR technique, defined in the LoRaWAN network protocol, is a simple technique that allows an adjustment of transmission rate by reading the SNR (Signal-to-Noise Ratio) value. Due to the multiplicity and dynamics of IoT scenarios, it is necessary to investigate ADR techniques that establish a good compromise between coverage and capacity. This thesis aims to investigate IoT scenarios of livestock in semi-confinement, especially in time-varying scenarios (emergence of concentrated traffic demand, network with mobile sensors, for example). Preliminary results using the ns-3 simulator demonstrate the need to dynamically adapt the ADR parameters, as each scenario requires different ADR strategies (or different parameterization of pre-existing strategies). Finally, we propose an adaptation of classic ADR algorithms to promote flexibility between coverage and capacity in such scenarios.